all AI news
480B LLM as 128x4B MoE? WHY?
April 26, 2024, noon | code_your_own_AI
code_your_own_AI www.youtube.com
Short introduction to MoE and then a comparison between different model architectures, followed up by a causal reasoning test (following test suite published by Stanford Univ).
Can a relatively small LLM, with below eg 5 Billion free trainable parameters, solve complex reasoning tasks. We evaluated this in my last video on PHI-3 MINI. …
architecture architectures causal comparison designs expert insights introduction llm llms moe reasoning snowflake stanford tasks tech tech insights test unique
More from www.youtube.com / code_your_own_AI
Understand DSPy: Programming AI Pipelines
1 day, 11 hours ago |
www.youtube.com
Latest Insights in AI Performance Models
3 days, 10 hours ago |
www.youtube.com
New Discovery: Retrieval Heads for Long Context
5 days, 11 hours ago |
www.youtube.com
Multi-Token Prediction (forget next token LLM?)
6 days, 10 hours ago |
www.youtube.com
LLMs: Rewriting Our Tomorrow (plus code) #ai
1 week, 1 day ago |
www.youtube.com
Autonomous AI Agents: 14 % MAX Performance
1 week, 3 days ago |
www.youtube.com
480B LLM as 128x4B MoE? WHY?
1 week, 5 days ago |
www.youtube.com
No more Fine-Tuning: Unsupervised ICL+
1 week, 6 days ago |
www.youtube.com
Jobs in AI, ML, Big Data
Lead Developer (AI)
@ Cere Network | San Francisco, US
Research Engineer
@ Allora Labs | Remote
Ecosystem Manager
@ Allora Labs | Remote
Founding AI Engineer, Agents
@ Occam AI | New York
AI Engineer Intern, Agents
@ Occam AI | US
AI Research Scientist
@ Vara | Berlin, Germany and Remote